In 2009, they decided to shift base to India because they saw the scope to help a larger pool of cancer patients.
Traditionally, there are a couple of methods doctors use to select a cancer drug.
One uses bio-markers to track the effects of cancer treatment. Biomarkers can be specific cells, molecules, or genes, gene products, enzymes, or hormones which can determine which cancer drug isn't effective — but it can't tell which drug works for a specific patient.
Another method, the Xenograft mouse model, transplants cancer cells from humans into mice. Mice are then dosed with different drugs, until the right one is found. It's a slow method though. According to Dr. Majumder, it can take four to six months for cancer cells to grow inside a mouse, which delays the drug selection process.
For different reasons, neither of these methods are as effective as people hope. The founders of Mitra Biotech knew this, and worked on a solution to theproblem. That solution is called the CANScriptTM model.
The CANScriptTM model is a platform that hosts cancer tumors in its native form in a specially customized incubator and provides a micro-environment where different drugs can be tried on cancer tumors. The CANScriptTM model shrinks the time it takes to select the most appropriate drug from months to just seven days.
How Top CIOs Tackle Big Data, Analytics and Cloud Security
Part of the reason the CANScriptTM model can do that is thanks to its robust analytical framework. CANScriptTM results are expressed in a predictive measure called the "M-Score". The CANScriptTM model measures the effect of multiple drugs on a patient's tumor using as many as 17 different parameters to assess the key effects on cell viability, cell death, tumor morphology, and cell proliferation. The raw data from these assays are converted into numerical scores that range from 0-100. Mitra's proprietary algorithm then assigns appropriate weightages to individual assays and then converts them to a single, numerical, predictive score called M-Score. The higher the M-Score, the greater the probability that a given drug combination works for a specific patient.
The M-Score algorithm was developed by simultaneously measuring patients' clinical responses to given drug combinations as well as CANScriptTM measurements across multiple dimension (17 assays measuring the tumor's response to drugs). Advanced bio-informatic tools were used to assign differential weightages and arrive at a unique algorithm with the least cumulative penalty when comparing the prediction to actual clinical outcome in 1,000 patients in the training set. This was further confirmed in a separate set of 1,000 patients in the test cohort. Cumulatively, CANScriptTM has a very high sensitivity and specificity for different class of drugs-chemotherapeutics, biologicals, and targeted drugs. Data analytics is the key to the spectacular success of CANScriptTM model.
Sign up for CIO Asia eNewsletters.